From owner-freebsd-questions@FreeBSD.ORG Tue May 4 15:49:45 2004 Return-Path: Delivered-To: freebsd-questions@freebsd.org Received: from mx1.FreeBSD.org (mx1.freebsd.org [216.136.204.125]) by hub.freebsd.org (Postfix) with ESMTP id 7780E16A4CE for ; Tue, 4 May 2004 15:49:45 -0700 (PDT) Received: from svaha.com (svaha.com [38.113.6.50]) by mx1.FreeBSD.org (Postfix) with ESMTP id 2FE3643D48 for ; Tue, 4 May 2004 15:49:44 -0700 (PDT) (envelope-from meconlen@obfuscated.net) Received: from [10.140.1.78] (noc.neutelligent.com [64.156.25.3]) (AUTH: LOGIN meconlen) by svaha.com with esmtp; Tue, 04 May 2004 18:49:40 -0400 Mime-Version: 1.0 (Apple Message framework v613) Content-Transfer-Encoding: 7bit Message-Id: <527A4BE3-9E1D-11D8-BD41-00039367611E@obfuscated.net> Content-Type: text/plain; charset=US-ASCII; format=flowed To: freebsd-questions@freebsd.org From: Michael Conlen Date: Tue, 4 May 2004 18:49:38 -0400 X-Mailer: Apple Mail (2.613) Subject: Disk Usage X-BeenThere: freebsd-questions@freebsd.org X-Mailman-Version: 2.1.1 Precedence: list List-Id: User questions List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Tue, 04 May 2004 22:49:45 -0000 I have a NFS server running FreeBSD-4.9-RELEASE. It's run fine for several months with five FreeBSD 4.9 systems mounting it's filesystems. Suddenly something started using disk space at the rate of 10 GB/hour on one of the filesystems (which has exported directories). The catch is that a du -k shows a total usage for that file system of much less than df -k. du -k essentially shows the disk usage before the available space started to disappear! Normally I'd presume someone's hiding files under a mount point when I see this but nothings mounted on a directory in this filesystem. Upon reboot the space is not used anymore, df -k and du -k report similar values. Quite simply odd. Some other details... ...this has happened twice in one day, and the rate of "ghost" disk usage is constant and identical in both graphs. The file server is used to serve files to clustered web servers. There's considerable write activity to the NFS server all the time (40-60Mbit/sec) and moderate read access (~10Mbit/sec). Any ideas what would cause the df -k and du -k discrepancy? -- Michael Conlen